how good is the data reliability with 90s and 2000s media?

[legacy]

I don't know if it's the right area of the forum, I'd like to discuss about CDROM, DVDRAM, IomegaZip, MO-disks, tapes etc, and their data reliability

a couple of days ago I was shocked by a couple of videos on Youtube where guys pointed out their negative experiences with CD ROMs as media for their own backup. They complained their data completely lost after 5 years of storage in CD ROMs, pointing out that their CDs were perfectly conserved and kept clean without scratches, but all the data is gone lost since the media is unreadable.

This is what they said in the video: what is your experience?

I'd like to understand Magneto-optical drive, whose technology should be better and more reliable, since the disc consists of a ferromagnetic material sealed beneath a plastic coating, and the only physical contact is during recording when a magnetic head is brought into contact with the side of the disc opposite to the laser. During reading, a laser projects a beam on the disk and, according to the magnetic state of the surface, the reflected light varies due to the magneto-optic Kerr effect. During recording, the laser power is increased to heat the material to the Curie point in a single spot. This enables an
electromagnet positioned on the opposite side of the disc, to change the local magnetic polarization. The polarization is retained after the temperature drops. Each write cycle requires both a pass to erase
a region, and another pass to write information. Both passes use the laser to heat the recording layer; the magnetic field is used for actually changing the magnetic orientation of the recording layer. The electromagnet reverses polarity for writing, and the laser is pulsed to record spots of "1" over the erased region of "0". As a result of this two-pass process, it takes twice as long to write data as it does
to read it.

I'd like to understand * what * in this process is subject to degrade. It looks the laser diode seems to be the most fragile element in the device, but it can be replaced.

what is your experience with MO-disks? And/or with tape as DAT, new DAT (DDS4), DLT and LTO?


Say, about me, I don't usually need Terabyte, 4.5Gbyte of media is more than OK since I usually need to archive projects, composed by C/C++/Ada sources, followed by tons of pdf and EagleCAD and gEDA sources.



I am not willing to use just one kind of media for my backups, rather I am using different kinds of media, including SSD, SD-cards, USB-pen drives and clouds, putting the same data on different media. This increases my chance my data will last longer.

[rsjsouza]

I threw away all my Zip 100MB disks a long time ago, but I have 23 year-old CD-Rs still fully readable and 28 year-old 3-1/2in floppy disks mostly readable.

I also have a few 20 year-old 5-1/4in floppy disks in readable state. They were not stored in temperature and humidity controlled environments, but they were surely kept away from heat and sunlight.

Also, I have never lived by the sea but my father had. His CD-Rs were still ok, but all his floppy disks were gone.

[tooki]

CD-R depends on the quality of blank discs (earlier ones were better than later, cheaper ones), the drive used to burn it, and the drive used to read it. I've actually had pretty good luck with CD-R, but then again I always bought premium discs, and always kept them in binders that keep them away from any light.

I have yet to have a Zip disk fail. Mine from the 90s still work fine. Same with floppies, as best I can tell. Magnetic media in general fares well, provided it was of decent quality and wasn't stored in awful conditions.

MO disks are likely the most reliable of all recordable media. Unlike every other recordable media, which use a single method (optical or magnetic) to both write and read (meaning that light or magnetic fields alone are enough to potentially damage data), MO uses optical to read, but both optical and magnetic to write, as you described. (Neither optical nor magnetic alone can write  to them — or damage the data.) There are a few variations in the actual write process, but they all rely on the combination of laser and magnetics to write, and laser to read. For example, MiniDisc does single-pass writing (including overwriting) by using a continuous laser to heat the material to the curie point, and then using the magnetic head to write the data. But is there any MO still in production these days??


As for me, I've long since switched to just keeping everything on hard disks, with automatic daily backups (Time Machine). If/when a disk starts to go bad, I replace aggressively, and make sure the backups are working. Haven't lost any data, ever, to media failure. (Just once, due to human error.)

[rdl]

It's been over five years since I burned a disk. I keep all my data on hard drives I use mirrored pairs with duplicates. They're on two different machines. In a couple of years I'll probably add some new drives and retire the oldest.

[David Hess]

I have yet to have a MO disk fail but new drives and media are difficult to find.  I will probably run out of MO drives before I run out of media.  Of the various options, I think these are the most reliable except for the lack of new hardware.

5-1/4" floppy disks were more reliable than 3-1/2 inch disks.  Especially there at the end, 3-1/2 inch drives and media were pretty poor.

CD media reliability depends on the manufacturer and CD writer.  The same goes for DVR+R media which should be tougher because the write layer is sandwiched between two disks.  I have only had failures with cheap media.  Use a writer which supports reading back the disk and reporting the various correctable errors to verify that the drive and media are working correctly and use the slowest writing speed.

ZIP drives depends on an embedded magnetic servo on the disk so the disks are very fragile.  Floptical type drives like the LS-120 were much more reliable.

NAND Flash is terrible due to low retention time.  In theory a NAND Flash drive could be kept powered and scrubbed.  SSDs which have power loss protection should be doing this automatically when powered.

[precaud]

In 38+ years of computing I have had one hard drive failure. A Connor SCSI1 drive, I believe it was. I stopped backing up years ago. The most I do is keep duplicates of important files on separate computers.

[David Hess]

My shelf of dead hard drives currently includes 5 Maxtors, 1 WD, and 1 Quantum.  I also had a group of Seagates (7200.10s? 7200.11s?) get returned a couple weeks after buying them and Seagate's customer support personnel were jerks so I no longer deal with them for anything.

[james_s]

I've had about 10 hard drives fail in various things, some people swear by one brand or another but every brand seems to have had a bad series now and then and most have made some really good drives too.

I have a bunch of CD-Rs from 15-20 years ago and haven't had one become unreadable yet. I've tended to use good quality blanks though.

[tooki]

5-1/4" floppy disks were more reliable than 3-1/2 inch disks.  Especially there at the end, 3-1/2 inch drives and media were pretty poor.

I think it's more about when the disks were made than what size. Early 3.5" were totally fine, too. It's the late-90s ones that suck. I've never had a quality 3.5" disk fail, as far as I can remember. But I always used quality disks, like 3M/Imation and BASF.

CD media reliability depends on the manufacturer and CD writer.  The same goes for DVR+R media which should be tougher because the write layer is sandwiched between two disks.  I have only had failures with cheap media.  Use a writer which supports reading back the disk and reporting the various correctable errors to verify that the drive and media are working correctly and use the slowest writing speed.

Correctable errors? That's not how CD-R/DVD-R work…

As for speed: No, the burn speed needs to be matched to the media. Using too fast a burn speed reduces reliability, but so does too slow a burn speed!

ZIP drives depends on an embedded magnetic servo on the disk so the disks are very fragile.  Floptical type drives like the LS-120 were much more reliable.

There's nothing about an embedded servo that makes a disk fragile. That's a totally specious argument. Hard disks rely on embedded servo, too, you know — and at tolerances FAR tighter than in a Zip drive. Other than the click of death, which is one very specific failure mode, Zip has proven to be very reliable. There's no evidence that LS120 is any more or less reliable than Zip, but its far smaller adoption makes it hard to really know its failure rate.

My shelf of dead hard drives currently includes 5 Maxtors, 1 WD, and 1 Quantum.  I also had a group of Seagates (7200.10s? 7200.11s?) get returned a couple weeks after buying them and Seagate's customer support personnel were jerks so I no longer deal with them for anything.

I've got a bunch of failed Seagates, a WD, a Hitachi and various old Quantums and Connors in vintage computers.

In 38+ years of computing I have had one hard drive failure. A Connor SCSI1 drive, I believe it was. I stopped backing up years ago. The most I do is keep duplicates of important files on separate computers.

Not backing up is really dumb. Even if you've had amazing luck with hard drives, drive failure is but one failure mode. Human error is a far larger one, as is software failure. That said, it sounds as though you ARE backing up, just selectively.

I've had about 10 hard drives fail in various things, some people swear by one brand or another but every brand seems to have had a bad series now and then and most have made some really good drives too.

Agreed. No one brand has made it without having some models that sucked. Seagate has served me the worst overall, Hitachi the best (but they're too expensive now, oftentimes), and WD has been my primary brand for about 10 years now. (Whereas in the late 90s-early 00s I wouldn't touch WD with a stick.)

[Circlotron]

I’ve had 720kB 3-1/2” disks fail.
The fact that I drilled a hole in the corner to make it 1.44MB clearly had nothing to do with it. 

[David Hess]

Correctable errors? That's not how CD-R/DVD-R work…

That is exactly how they work; CDs and DVDs always have an assortment of correctable errors and so do hard drives.  (1)  I do not remember the details for data CDs but for DVDs, there are two layers of error correction with the second layer catching errors that the first layer missed.  Good DVD burners (2) have the capability of doing a surface scan to return the number of errors corrected by the first and second layers of error correction and this reveals the general quality of the burned disk.  (3) Only uncorrectable errors show up in the typical media verify function.

(1) A major reason hard drives have moved to using larger sector sizes is because error correction becomes less efficient as the physical sector length becomes shorter (greater linear density) compared to the defect density.  I suspect the same thing was responsible for larger sector sizes in higher density MO disks except where compatibility was required.  DVDs had this covered from the start with larger sector sizes and a combination of short and long error correction to handle a much more hostile environment.
(2) Plextor burners were good for this until they stopped making their own drives.
(3) Which is also handy for evaluating the best write speed.  I never found a drive and media combination which produced fewer correctable errors at higher write speeds.

ZIP drives depends on an embedded magnetic servo on the disk so the disks are very fragile.  Floptical type drives like the LS-120 were much more reliable.

There's nothing about an embedded servo that makes a disk fragile. That's a totally specious argument. Hard disks rely on embedded servo, too, you know — and at tolerances FAR tighter than in a Zip drive.

It is not the embedded servo itself but the embedded servo in combination with the low coercivity of Zip media.  This makes it possible to easily erase the magnetic servo with an external magnetic field at which point the disk is trashed because the drive has no way to reformat it.  Hard drives which all use an embedded servo these days have a media coercivity which is so high that practically only the hard drive write head can affect it.  It takes a massive degausser to erase a modern hard drive.

I have always wondered if the loss of the embedded servo tracks on Zip media was responsible for damaging the drive leading to the click of death but I avoided them and Jaz drives.  I decided using an embedded servo on soft media was bad idea from the start except for planned obsolescence.

The old Syquest drives were the same way but with a "hard" magnetic media, damaging the embedded servo was not likely; I never saw it happen but with a flying head, head crashes were always possible.

There's no evidence that LS120 is any more or less reliable than Zip, but its far smaller adoption makes it hard to really know its failure rate.

At least when I accidentally wipe an LS120 disk with a magnet, I can reformat it and use it.  I never used LS120 in any significant way either but only because it never became popular and I was already using 3-1/2 inch MO although I think burnable CDs and DVDs are even better for archival and backup purposes because you cannot make a mistake and erase your old data.

MO disks are the best as far as toughness since they are "hard" magnetic media except during the write process, use optical tracking so the embedded servo data is permanent, and are a non-contact media.  Unfortunately optics limit areal density compared to magnetic media.

[tooki]

Correctable errors? That's not how CD-R/DVD-R work…

That is exactly how they work; CDs and DVDs always have an assortment of correctable errors and so do hard drives.

Ohhh, derp, I somehow thought you meant "correctable" in the sense of going back and rewriting a failed sector or something…

(1)  I do not remember the details for data CDs but for DVDs, there are two layers of error correction with the second layer catching errors that the first layer missed.  Good DVD burners (2) have the capability of doing a surface scan to return the number of errors corrected by the first and second layers of error correction and this reveals the general quality of the burned disk.  (3) Only uncorrectable errors show up in the typical media verify function.

CD-ROMs use eight-to-fourteen modulation as one layer of redundancy, the other is the cross-interleaved Reed-Solomon encoding, just like an audio CD. But unlike an audio CD, CD-ROMs also add another layer of Reed-Solomon encoding and a CRC. (That's why a 74-minute blank CD holds 74 minutes of audio, which is actually about 750MB of audio data, will only hold 650MB as a CD-ROM.)

(1) A major reason hard drives have moved to using larger sector sizes is because error correction becomes less efficient as the physical sector length becomes shorter (greater linear density) compared to the defect density.  I suspect the same thing was responsible for larger sector sizes in higher density MO disks except where compatibility was required.  DVDs had this covered from the start with larger sector sizes and a combination of short and long error correction to handle a much more hostile environment.

Yeah, I remember reading that!

(2) Plextor burners were good for this until they stopped making their own drives.

Yes, it's a pity Plextor and Yamaha stopped making drives — those two were hands-down the best!

ZIP drives depends on an embedded magnetic servo on the disk so the disks are very fragile.  Floptical type drives like the LS-120 were much more reliable.

There's nothing about an embedded servo that makes a disk fragile. That's a totally specious argument. Hard disks rely on embedded servo, too, you know — and at tolerances FAR tighter than in a Zip drive.

It is not the embedded servo itself but the embedded servo in combination with the low coercivity of Zip media.  This makes it possible to easily erase the magnetic servo with an external magnetic field at which point the disk is trashed because the drive has no way to reformat it.

In theory, yes. But I've literally never heard of that happening. Not once. Remember that even floppy disks aren't nearly as sensitive to magnetic fields as people think, and Zip media would have to have far lower coercivity than those.

I have always wondered if the loss of the embedded servo tracks on Zip media was responsible for damaging the drive leading to the click of death but I avoided them and Jaz drives.

The click of death is a bit of a red herring IMHO, insofar as it actually describes a number of different failure modes. (And as a computer technician right when these things were popular, I think the failure rate is WILDLY exaggerated.)

In the Zip, one "click" was when misaligned heads caused trouble reading the servo tracks, causing seek attempts that sound clicky. It is not clear to me whether head misalignment was contagious, insofar as I don't know of any way that a bad drive could damage a disk in a way that that disk can damage further drives.

Another Zip "click" was catastrophic head damage, which would cause edge damage to the disk, which in turn would rip the heads off the next drive it was inserted in, damaging the drive and causing it to cause damage to every disk inserted, in a vicious, contagious cycle.

The Jaz was a totally different technology (removable hard disk, same as SyQuest and Orb). Its click of death was basically "standard" hard disk seek errors, plus dust ingress, dust created by shutter wear, etc.

I decided using an embedded servo on soft media was bad idea from the start except for planned obsolescence.

I think that was an unfounded fear. (As for planned obsolescence… I don't really believe in that. IT has more than enough inherent obsolescence, such that adding artificial obsolescence is absolutely unnecessary!)

The old Syquest drives were the same way but with a "hard" magnetic media, damaging the embedded servo was not likely; I never saw it happen but with a flying head, head crashes were always possible.

As I said about Jaz, they sucked in other ways. (FWIW, I saw Jaz drives fail at a rate of probably 10:1 vs Zip. The failure rate on Jaz was absolutely shocking, and I think THAT is what actually led to Zip's undeservedly bad reputation.)

There's no evidence that LS120 is any more or less reliable than Zip, but its far smaller adoption makes it hard to really know its failure rate.

At least when I accidentally wipe an LS120 disk with a magnet, I can reformat it and use it.

For sure, a degaussed Zip can't be reused, but I've never actually heard of it happening. As I said, even floppy disks aren't that sensitive to magnets. It takes an oscillating magnetic field to affect them.

I never used LS120 in any significant way either but only because it never became popular and I was already using 3-1/2 inch MO

MO is something I wish had caught on more, too.

although I think burnable CDs and DVDs are even better for archival and backup purposes because you cannot make a mistake and erase your old data.

True. But the inherent suckiness of media that's written with the same light that reads it (or ambient light, if stored poorly!) makes them questionable in the long run — as experience is showing.

MO disks are the best as far as toughness since they are "hard" magnetic media except during the write process, use optical tracking so the embedded servo data is permanent, and are a non-contact media.  Unfortunately optics limit areal density compared to magnetic media.

Yep. At least in MiniDisc, it uses a pregroove, much like CD-R. Data MD is something I REALLY wish Sony had developed further. It was such a convenient format, and MiniDisc media has proven to be spectacularly reliable in the long run. (But the Data MD drives Sony made were much, much, much too slow.)

[David Hess]

The Jaz was a totally different technology (removable hard disk, same as SyQuest and Orb). Its click of death was basically "standard" hard disk seek errors, plus dust ingress, dust created by shutter wear, etc.

I thought Jaz used a flexible Bernoulli disk.  SyQuest used a hard media.  Zip was a floppy technology with a head that contacts the film disk but with an embedded servo.

There's no evidence that LS120 is any more or less reliable than Zip, but its far smaller adoption makes it hard to really know its failure rate.

At least when I accidentally wipe an LS120 disk with a magnet, I can reformat it and use it.

For sure, a degaussed Zip can't be reused, but I've never actually heard of it happening. As I said, even floppy disks aren't that sensitive to magnets. It takes an oscillating magnetic field to affect them.

I saw a lot of magnetically erased Zip disks in connection with drives which had suffered the click of death.  I only encountered a few floppy disks which had been magnetically erased but I suspect people were more aware and careful with them.

It definitely did not take much of a magnet to degauss a floppy diskette.  I tested this several times and the swipes could be seen with a good surface scan utility.

I never used LS120 in any significant way either but only because it never became popular and I was already using 3-1/2 inch MO

MO is something I wish had caught on more, too.

Apparently MO was more popular outside the US.  Sony's MiniDisc was of course popular in Japan.

MO disks are the best as far as toughness since they are "hard" magnetic media except during the write process, use optical tracking so the embedded servo data is permanent, and are a non-contact media.  Unfortunately optics limit areal density compared to magnetic media.

Yep. At least in MiniDisc, it uses a pregroove, much like CD-R. Data MD is something I REALLY wish Sony had developed further. It was such a convenient format, and MiniDisc media has proven to be spectacularly reliable in the long run. (But the Data MD drives Sony made were much, much, much too slow.)

I was aware of Sony's MiniDisc but even back then I was avoiding everything Sony touched.  My MO drives were all made by Fujitsu but of course get practically no use now; I have them installed in my legacy systems along with my IDE Zip, LS120, and DVD-RAM drives which I inherited.  The DVD-RAM drive is great for recovering damaged CDs and DVDs.

[tooki]

The Jaz was a totally different technology (removable hard disk, same as SyQuest and Orb). Its click of death was basically "standard" hard disk seek errors, plus dust ingress, dust created by shutter wear, etc.

I thought Jaz used a flexible Bernoulli disk.  SyQuest used a hard media.  Zip was a floppy technology with a head that contacts the film disk but with an embedded servo.

Nope. As I said, people really confounded the two. Jaz is simply removable hard disk. Slide open the shutter and you see the platter inside.

I saw a lot of magnetically erased Zip disks in connection with drives which had suffered the click of death.  I only encountered a few floppy disks which had been magnetically erased but I suspect people were more aware and careful with them.

Do you know they were magnetically erased? (Can this even be absolutely determined with a consumer drive?) Or had they just been written to by heads so misaligned that other drives could not read it? The only thing I could possibly envision is heads so misaligned that they overwrite the servo track.

It definitely did not take much of a magnet to degauss a floppy diskette.  I tested this several times and the swipes could be seen with a good surface scan utility.

And yet I remember seeing or reading about actual tests performed, and even things like sticking disks to a fridge with a magnet didn't affect them.

Apparently MO was more popular outside the US.  Sony's MiniDisc was of course popular in Japan.

For sure, MO was popular among graphic designers in Europe, though to some extent it was displaced by Zip and Jaz, which were much faster.

As for MiniDisc, a more accurate characterization is that (among developed nations) it was a failure in North America. It was quite successful in Europe, not just Japan. (Because English-language tech media is so US-heavy, we really get the US-centric perception that MD was a failure except in Japan, but that's really quite untrue.)
 

I was aware of Sony's MiniDisc but even back then I was avoiding everything Sony touched.

How come? MD was released while Sony was still in its heyday.

My MO drives were all made by Fujitsu but of course get practically no use now; I have them installed in my legacy systems along with my IDE Zip, LS120, and DVD-RAM drives which I inherited.  The DVD-RAM drive is great for recovering damaged CDs and DVDs.

What's so funny is that most people have no idea that a great many DVD burners support DVD-RAM, too — you just have to take the disc out of the DVD-RAM shell!

[David Hess]

I saw a lot of magnetically erased Zip disks in connection with drives which had suffered the click of death.  I only encountered a few floppy disks which had been magnetically erased but I suspect people were more aware and careful with them.

Do you know they were magnetically erased? (Can this even be absolutely determined with a consumer drive?) Or had they just been written to by heads so misaligned that other drives could not read it? The only thing I could possibly envision is heads so misaligned that they overwrite the servo track.

They could not be formatted indicating a loss of the embedded servo track.  I never had a drive to test with at the time but I relayed my concerns to others who did and they confirmed it.

It definitely did not take much of a magnet to degauss a floppy diskette.  I tested this several times and the swipes could be seen with a good surface scan utility.

And yet I remember seeing or reading about actual tests performed, and even things like sticking disks to a fridge with a magnet didn't affect them.

Refrigerator magnets are incredibly weak.  I tested using a small horseshoe magnet which is more typical of a useful magnet like a common magnetic parts wand.

I was aware of Sony's MiniDisc but even back then I was avoiding everything Sony touched.

How come? MD was released while Sony was still in its heyday.

The DRM fiasco with MiniDisc audio was enough to discern that Sony had passed their expiration date.  The release of their proprietary Memory Stick a couple years later and audio CD root kit reinforced this.  Sony is not to be trusted with anything and they can burn in hell as far as I am concerned.

[SeanB]

I had a set of old Verbatim CDR disks that were backups, and I tried a few of them to see if they still worked, as they were written by a HP writer that was a parallel port drive, and ran with Win95. A whole 2x write speed, and no buffer aside from the 16k on the drive, so it was very prone to making coasters, and the disks were around $3 each, and even more if they were the rewritable ones. CDR was the month end backup, and the CDRW was the daily one, rotated on a 2 week cycle, in addition to the QIC drive backup.

After 20 years of being in storage all the disks were readable, so they had a 30 second data erasure cycle in the microwave, making a nice light show and a smell of well toasted acrylic. The later ones that were not Verbatim, and which were in 50 packs, were almost all not readable, or had many bad sector reads on reading. CDRW disks though, despite having had a good number of erase cycles on them, were still fine, even the scratched ones. Still have a part pack of the Verbatim disks around, as there was a migration to DAT tape, and the HP writer was retired.

[tooki]

I saw a lot of magnetically erased Zip disks in connection with drives which had suffered the click of death.  I only encountered a few floppy disks which had been magnetically erased but I suspect people were more aware and careful with them.

Do you know they were magnetically erased? (Can this even be absolutely determined with a consumer drive?) Or had they just been written to by heads so misaligned that other drives could not read it? The only thing I could possibly envision is heads so misaligned that they overwrite the servo track.

They could not be formatted indicating a loss of the embedded servo track.  I never had a drive to test with at the time but I relayed my concerns to others who did and they confirmed it.

So yeah, borked servo tracks, but no way of knowing whether it was bulk degaussing or overwriting by a misaligned head.

It definitely did not take much of a magnet to degauss a floppy diskette.  I tested this several times and the swipes could be seen with a good surface scan utility.

And yet I remember seeing or reading about actual tests performed, and even things like sticking disks to a fridge with a magnet didn't affect them.

Refrigerator magnets are incredibly weak.  I tested using a small horseshoe magnet which is more typical of a useful magnet like a common magnetic parts wand.

It's been ages, but I'm pretty sure the test I saw used strong magnets. The fridge test was just one of the things they tried.

I was aware of Sony's MiniDisc but even back then I was avoiding everything Sony touched.

How come? MD was released while Sony was still in its heyday.

The DRM fiasco with MiniDisc audio was enough to discern that Sony had passed their expiration date.  The release of their proprietary Memory Stick a couple years later and audio CD root kit reinforced this.  Sony is not to be trusted with anything and they can burn in hell as far as I am concerned.

What a oversimplification and mischaracterization of the situation! :O

As with most things, the reality is far more nuanced.

Sony added copy protection to MiniDisc because it was forced to by the RIAA. (So was Philips with DCC, and Sony with DAT prior, and the whole industry with audio CD recorders after that.) Not that I'd call it a "fiasco" since it actually worked really well in practice. The only practical limitation was the inability to make a second-generation digital copy. No limits at all on analog copies or first-generation digital copies. Given that MD's primary use case was basically "mix tapes" copied from CDs (not MD-to-MD copies!), it posed zero hassle in everyday use. (It was nothing like modern DRM that uses encryption, with fickle handshaking, slow VMs, etc… HDMI and blu-ray, I'm looking at you!)

That you see Sony as being "beyond its expiration" because of this is indicative only of the fact that Sony was one of the few companies even bothering to create new formats at all! ANY vendor creating new formats at that time was subject to the lawsuits that resulted in the DRM. Remember, we are talking late 80s here, and the RIAA had NO idea how to deal with the potential of high-quality digital copies, so they dug in their heels and called their lawyers.

In the end, Sony nearly succumbed to corporate schizophrenia resulting from its purchase of Columbia (i.e. CBS Records and Columbia Pictures), culminating in Sony suing itself in 2000. (Basically, the RIAA, which Sony Music is a member of, sued a music service which was partly owned by Sony Corporation, the electronics arm.) So you had Sony Corporation engineering amazing technology, and then being convinced to cripple it by the entertainment arm. What had been intended to provide a guaranteed source of prerecorded material for new formats (to give them a head start) ended up putting Sony on both sides of the same legal battle!!  And it's Sony Music, not Sony Corporation, that released that horrific DRM. By the early 2000s, Sony Corporation was so worn down that they basically let the music industry dictate the terms, hence the insane DRM used on Sony's early MP3 players (the Network Walkman bullcrap, and the MP3-copying software for the later MiniDisc units).

DRM in the music industry only got worse and worse until Apple really fought it by making the iTunes music store DRM-free. (I mean, the writing had been on the wall for years, but without a corporate giant to fight it, we might still have been stuck with the RIAA's demands.)


I think it's also important to recognize that Sony is a keiretsu, one of the giant Japanese conglomerates with disparate industries, loosely organized into a group. The members of the group have very little day-to-day involvement with the runnings of the other members, since they are separate companies. So IMHO it's rather unfair to condemn and disparage Sony Corporation for the actions of Sony Music — they're no more connected to one another than Nikon (the camera company) and Kirin (the brewery) are to each other, in that they're both member companies of the Mitsubishi keiretsu.



The Memory Stick, by the way, was not a "few" years after MiniDisc, it was actually a solid 6 years later. It was during a time when tons of companies were introducing competing memory card formats, and like all the others, Sony licensed it to many other companies to try and make it an industry standard. It was also the very first with DRM support, which is after all why the SD format was later created. (Secure Digital is the DRM-capable version of the older MMC format from SanDisk.) In essence, SD began life as a proprietary format, but which was later transferred to an industry group.

[tooki]

After 20 years of being in storage all the disks were readable, so they had a 30 second data erasure cycle in the microwave, making a nice light show and a smell of well toasted acrylic. The later ones that were not Verbatim, and which were in 50 packs, were almost all not readable, or had many bad sector reads on reading. CDRW disks though, despite having had a good number of erase cycles on them, were still fine, even the scratched ones.

CD-RW is a phase-change medium using entirely different technology from CD-R, which is why a CD reader must be specifically designed to support reading CD-RW. (A CD-R uses dyes to mimic the optical properties of a pressed CD, the intent being that it be readable by ordinary CD readers.) The phase-change material is far more durable than the dyes in CD-Rs, hence the vast difference in long-term durability.

[David Hess]

The only writable CDs or DVDs that I have had which regularly started out bad were from RiData.  Running the correctable errors test I described showed that they were very marginal and they became unreadable within weeks to months.

Except for those and some other unknowns, all of the disks I have burned going back 20 years are readable.  Occasionally I find one of uncertain heritage with a couple of errors which I recover using IsoBuster and PAR sets.


Sony does not get points for going along with what their contemporaries were doing as well; this is not a game of who is better than Hitler.  Their later decisions show that they were all about futile gestures in an attempt to extract rents; they were willing partners with the RIAA.

When Sony's MiniDisc came out, I and my friends were already digitally ripping our CD collections to MP3s (Was that on 486s and Pentiums?  Get off my lawn!  *shakes fist*) so we had zero interest in their deliberately crippled products.  Apple's DRM free offerings were years too late for us.

DRM in the music industry only got worse and worse until Apple really fought it by making the iTunes music store DRM-free. (I mean, the writing had been on the wall for years, but without a corporate giant to fight it, we might still have been stuck with the RIAA's demands.)

My contemporaries were never stuck with anything and Apple had no effect on us.  The writing was on the wall because of us and we knew it.  As far as we were concerned, the RIAA (and Sony) lost that war before they recognized it as such.

[tooki]

When Sony's MiniDisc came out, I and my friends were already digitally ripping our CD collections to MP3s (Was that on 486s and Pentiums?  Get off my lawn!  *shakes fist*) so we had zero interest in their deliberately crippled products.

No, you weren't: MiniDisc was released in 1992 — this is why I think you're being much too critical. It was a completely different environment. If the MiniDisc had been released in 1999, which is kinda around where your memory is placing it, your criticism would be justified. But you're way, way out. When MD came out, there was nothing like it on the market. CD-R cost thousands and thousands for a burner and used $30 blanks. Tape was analog and not random-access like MD. MP3 was still being conceived. At release, MD was competing with tape, and it did that beautifully, even if the first-gen MD devices kinda sucked (the second gen a year or so later fixed all the shortcomings).

The MP3 algorithm wasn't published until 1993, and the first software decoder came in 1994, but CPUs at the time couldn't rip or play it in realtime — not that you'd have had the hard disk space for files anyway. In 1993, a typical hard disk was around 200MB. It didn't really matter, since the initial MP3 encoders were terrible so the music sounded awful. The very first realtime MP3 playback app was released in Fall 1995, but the seminal MP3 player app, Winamp, wasn't released until 1997. You'd have been on a Pentium II (or a PowerPC 604e on a Mac) by then. But in 1997, it was a niche thing still. It was over the next year or two that MP3s really exploded. Napster, which was instrumental in MP3s popularity, wasn't released until mid-1999.

[nctnico]

CD-R depends on the quality of blank discs (earlier ones were better than later, cheaper ones), the drive used to burn it, and the drive used to read it. I've actually had pretty good luck with CD-R, but then again I always bought premium discs, and always kept them in binders that keep them away from any light.

Same here. The trick is to use low writing speeds like 4x or 8x and good A-brand discs like Philips or Sony.

@tooki: good MP3 players which worked on a P1-133MHz where available in 1995 /1996. I know for sure because that is what the file date says on my first MP3s. And they don't sound awful. By that time hard drives started to get really cheap as well.

[james_s]

I remember when I first started ripping my CDs it took about 2x the length of the song to encode it. I don't recall what CPU I had at the time but I think it was a 100MHz Pentium. I remember being excited later on a faster machine when I could rip and encode in better than real time.

Minidisc was a great format, but that and DAT were both crippled by hopeless DRM.

[tooki]

CD-R depends on the quality of blank discs (earlier ones were better than later, cheaper ones), the drive used to burn it, and the drive used to read it. I've actually had pretty good luck with CD-R, but then again I always bought premium discs, and always kept them in binders that keep them away from any light.

Same here. The trick is to use low writing speeds like 4x or 8x and good A-brand discs like Philips or Sony.

The oldest burned discs I have at hand are TDK, Fujifilm and Imation with the dark azo dyes (so likely made by Taiyo Yuden). Later CD-Rs all moved to cheaper dyes that aren't as dark and aren't as good (cyanine and then phthalocyanine).

One other HUGE thing: I bought the 74min/650MB discs as long as I could. The wider track spacing improves reliability. It really pissed me off that 80min/700MB blanks took over, since the overwhelming majority of burned CDs never even hit 650MB, so nearly always, the extra space went to waste, but still took its reliability penalty. (I remember this because Yamaha did actual market research on this, and found that only a tiny percentage of discs are ever close to filled. So they made some burners that had a special option that burned the disc in a weird way to make the pits and lands longer than they should be, somehow tricking the reader into spinning the disc a bit faster to compensate. The longer pits meant reduced capacity but increased accuracy, and so theoretically better performance. I don't remember if this was available to data CDs, too, or only for audio.)

Most people have totally forgotten that there used to be 63min/550MB CD-R blanks, which presumably would have had better reliability still. But they disappeared by around 1999 I think.

@tooki: good MP3 players which worked on a P1-133MHz where available in 1995 /1996. I know for sure because that is what the file date says on my first MP3s. And they don't sound awful. By that time hard drives started to get really cheap as well.

By late 95 they were a lot better. The ones from 93-94 were pretty bad. I remember trying them as a test and being appalled.  But later MP3 encoders (like by 1998-99) were still noticeably better than the 96-97 ones. Less sparkling and stuff like that.

[legacy]

thanks for your answers, guys, very appreciated 

I am in the process of building a new SCSI media-tower loaded with
- DLT8000 tape (IBM tapes)
- DDS4 tape (IBM tapes)
- CD writer (Plextor writer, Verbatim premium dry-CDs)
- DVDRAM
- micro, mechanical removable hard drive (CF2 Microdrive)
- common 3.5" mechanical removable hard drive (SCA-80)

I am duplicating my backups in different kind of media in order to have more probability to have them preserved.

I have removed from the media-tower technology like NAND-Flash Compact Flash, I will replace them with MicroDrive (mechanical removable hard drive, on CF2 interface)

But, I still wonder if/not to buy a MO unit: it's a bit expensive, and media are hard to find.


Besides the media tower, I am also duplicating backups with this NAS-project (I am the author), that uses common sATA disks in RAID1 (mirroring) configuration.

Hope to do my best to preserve my sources (C, C++, Pascal, Ada, Modula2, Oberon, and EDA (OrCad, EagleCAD) files ), SGI-applications, and file repositories 

[legacy]

MO drivers and media are a bit costly: 500-600 euro for the driver and 50 euro for the media (4.5GB, re-writable). I am evaluating the purchase, but first I have to understand if it's worth with the effort.